orcs-oss/orcs-code
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add NVIDIA NIM and MiniMax provider support (#552)

Browse Source 
* feat: add NVIDIA NIM and MiniMax provider support

- Add nvidia-nim and minimax to --provider CLI flag
- Add model discovery for NVIDIA NIM (160+ models) and MiniMax
- Update /model picker to show provider-specific models
- Fix provider detection in startup banner
- Update .env.example with new provider options

Supported providers:
- NVIDIA NIM: https://integrate.api.nvidia.com/v1
- MiniMax: https://api.minimax.io/v1

* fix: resolve conflict in StartupScreen (keep NVIDIA/MiniMax + add Codex detection)

* fix: resolve providerProfile conflict (add imports from main, keep NVIDIA/MiniMax)

* fix: revert providerSecrets to match main (NVIDIA/MiniMax handled elsewhere)

* fix: add context window entries for NVIDIA NIM and new MiniMax models

* fix: use GLM-5 as NVIDIA NIM default and MiniMax-M2.5 for consistency

* fix: address remaining review items - add GLM/Kimi context entries, max output tokens, fix .env.example, revert to Nemotron default

* fix: filter NVIDIA NIM picker to chat/instruct models only, set provider-specific API keys from saved profiles

* chore: add more NVIDIA NIM context window entries for popular models

* fix: address remaining non-blocking items - fix base model, clear provider API keys on profile switch
This commit is contained in:
ArkhAngelLifeJiggy
2026-04-15 13:26:13 +01:00
committed by GitHub
parent 6b2121da12
commit 51191d6132
15 changed files with 628 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
.env.example
View File

@@ -225,6 +225,30 @@ ANTHROPIC_API_KEY=sk-ant-your-key-here
# GOOGLE_CLOUD_PROJECT=your-gcp-project-id # GOOGLE_CLOUD_PROJECT=your-gcp-project-id
# -----------------------------------------------------------------------------
# Option 9: NVIDIA NIM
# -----------------------------------------------------------------------------
# NVIDIA NIM provides hosted inference endpoints for NVIDIA models.
# Get your API key from https://build.nvidia.com/
#
# CLAUDE_CODE_USE_OPENAI=1
# NVIDIA_API_KEY=nvapi-your-key-here
# OPENAI_BASE_URL=https://integrate.api.nvidia.com/v1
# OPENAI_MODEL=nvidia/llama-3.1-nemotron-70b-instruct
# -----------------------------------------------------------------------------
# Option 10: MiniMax
# -----------------------------------------------------------------------------
# MiniMax API provides text generation models.
# Get your API key from https://platform.minimax.io/
#
# CLAUDE_CODE_USE_OPENAI=1
# MINIMAX_API_KEY=your-minimax-key-here
# OPENAI_BASE_URL=https://api.minimax.io/v1
# OPENAI_MODEL=MiniMax-M2.5
# ============================================================================= # =============================================================================
# OPTIONAL TUNING # OPTIONAL TUNING
# ============================================================================= # =============================================================================

10
src/components/ProviderManager.tsx
View File

@@ -1009,6 +1009,16 @@ export function ProviderManager({ mode, onDone }: Props): React.ReactNode {
label: 'Custom', label: 'Custom',
description: 'Any OpenAI-compatible provider', description: 'Any OpenAI-compatible provider',
}, },
{
value: 'nvidia-nim',
label: 'NVIDIA NIM',
description: 'NVIDIA NIM endpoint',
},
{
value: 'minimax',
label: 'MiniMax',
description: 'MiniMax API endpoint',
},
...(mode === 'first-run' ...(mode === 'first-run'
? [ ? [
{ {

31
src/components/StartupScreen.ts
View File

@@ -117,17 +117,28 @@ function detectProvider(): { name: string; model: string; baseUrl: string; isLoc
const baseUrl = resolvedRequest.baseUrl const baseUrl = resolvedRequest.baseUrl
const isLocal = isLocalProviderUrl(baseUrl) const isLocal = isLocalProviderUrl(baseUrl)
let name = 'OpenAI' let name = 'OpenAI'
// Override to Codex when resolved endpoint is Codex if (/nvidia/i.test(baseUrl) || /nvidia/i.test(rawModel) || process.env.NVIDIA_NIM)
if (resolvedRequest.transport === 'codex_responses' || baseUrl.includes('chatgpt.com/backend-api/codex')) { name = 'NVIDIA NIM'
else if (/minimax/i.test(baseUrl) || /minimax/i.test(rawModel) || process.env.MINIMAX_API_KEY)
name = 'MiniMax'
else if (resolvedRequest.transport === 'codex_responses' || baseUrl.includes('chatgpt.com/backend-api/codex'))
name = 'Codex' name = 'Codex'
} else if (/deepseek/i.test(baseUrl) || /deepseek/i.test(rawModel)) name = 'DeepSeek' else if (/deepseek/i.test(baseUrl) || /deepseek/i.test(rawModel))
else if (/openrouter/i.test(baseUrl)) name = 'OpenRouter' name = 'DeepSeek'
else if (/together/i.test(baseUrl)) name = 'Together AI' else if (/openrouter/i.test(baseUrl))
else if (/groq/i.test(baseUrl)) name = 'Groq' name = 'OpenRouter'
else if (/mistral/i.test(baseUrl) || /mistral/i.test(rawModel)) name = 'Mistral' else if (/together/i.test(baseUrl))
else if (/azure/i.test(baseUrl)) name = 'Azure OpenAI' name = 'Together AI'
else if (/llama/i.test(rawModel)) name = 'Meta Llama' else if (/groq/i.test(baseUrl))
else if (isLocal) name = getLocalOpenAICompatibleProviderLabel(baseUrl) name = 'Groq'
else if (/mistral/i.test(baseUrl) || /mistral/i.test(rawModel))
name = 'Mistral'
else if (/azure/i.test(baseUrl))
name = 'Azure OpenAI'
else if (/llama/i.test(rawModel))
name = 'Meta Llama'
else if (isLocal)
name = getLocalOpenAICompatibleProviderLabel(baseUrl)
// Resolve model alias to actual model name + reasoning effort // Resolve model alias to actual model name + reasoning effort
let displayModel = resolvedRequest.resolvedModel let displayModel = resolvedRequest.resolvedModel

8
src/services/api/openaiShim.ts
View File

@@ -1360,8 +1360,12 @@ class OpenAIShimMessages {
...filterAnthropicHeaders(options?.headers), ...filterAnthropicHeaders(options?.headers),
} }
const isGemini = isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI) const isGemini = isGeminiMode()
const apiKey = this.providerOverride?.apiKey ?? process.env.OPENAI_API_KEY ?? '' const isMiniMax = !!process.env.MINIMAX_API_KEY
const apiKey =
this.providerOverride?.apiKey ??
process.env.OPENAI_API_KEY ??
(isMiniMax ? process.env.MINIMAX_API_KEY : '')
// Detect Azure endpoints by hostname (not raw URL) to prevent bypass via // Detect Azure endpoints by hostname (not raw URL) to prevent bypass via
// path segments like https://evil.com/cognitiveservices.azure.com/ // path segments like https://evil.com/cognitiveservices.azure.com/
let isAzure = false let isAzure = false

22
src/utils/model/configs.ts
View File

@@ -37,6 +37,8 @@ export const CLAUDE_3_7_SONNET_CONFIG = {
gemini: 'gemini-2.0-flash', gemini: 'gemini-2.0-flash',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_3_5_V2_SONNET_CONFIG = { export const CLAUDE_3_5_V2_SONNET_CONFIG = {
@@ -48,6 +50,8 @@ export const CLAUDE_3_5_V2_SONNET_CONFIG = {
gemini: 'gemini-2.0-flash', gemini: 'gemini-2.0-flash',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_3_5_HAIKU_CONFIG = { export const CLAUDE_3_5_HAIKU_CONFIG = {
@@ -59,6 +63,8 @@ export const CLAUDE_3_5_HAIKU_CONFIG = {
gemini: 'gemini-2.0-flash-lite', gemini: 'gemini-2.0-flash-lite',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_HAIKU_4_5_CONFIG = { export const CLAUDE_HAIKU_4_5_CONFIG = {
@@ -70,6 +76,8 @@ export const CLAUDE_HAIKU_4_5_CONFIG = {
gemini: 'gemini-2.0-flash-lite', gemini: 'gemini-2.0-flash-lite',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_SONNET_4_CONFIG = { export const CLAUDE_SONNET_4_CONFIG = {
@@ -81,6 +89,8 @@ export const CLAUDE_SONNET_4_CONFIG = {
gemini: 'gemini-2.0-flash', gemini: 'gemini-2.0-flash',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_SONNET_4_5_CONFIG = { export const CLAUDE_SONNET_4_5_CONFIG = {
@@ -92,6 +102,8 @@ export const CLAUDE_SONNET_4_5_CONFIG = {
gemini: 'gemini-2.0-flash', gemini: 'gemini-2.0-flash',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_OPUS_4_CONFIG = { export const CLAUDE_OPUS_4_CONFIG = {
@@ -103,6 +115,8 @@ export const CLAUDE_OPUS_4_CONFIG = {
gemini: 'gemini-2.5-pro-preview-03-25', gemini: 'gemini-2.5-pro-preview-03-25',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_OPUS_4_1_CONFIG = { export const CLAUDE_OPUS_4_1_CONFIG = {
@@ -114,6 +128,8 @@ export const CLAUDE_OPUS_4_1_CONFIG = {
gemini: 'gemini-2.5-pro-preview-03-25', gemini: 'gemini-2.5-pro-preview-03-25',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_OPUS_4_5_CONFIG = { export const CLAUDE_OPUS_4_5_CONFIG = {
@@ -125,6 +141,8 @@ export const CLAUDE_OPUS_4_5_CONFIG = {
gemini: 'gemini-2.5-pro-preview-03-25', gemini: 'gemini-2.5-pro-preview-03-25',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_OPUS_4_6_CONFIG = { export const CLAUDE_OPUS_4_6_CONFIG = {
@@ -136,6 +154,8 @@ export const CLAUDE_OPUS_4_6_CONFIG = {
gemini: 'gemini-2.5-pro-preview-03-25', gemini: 'gemini-2.5-pro-preview-03-25',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
export const CLAUDE_SONNET_4_6_CONFIG = { export const CLAUDE_SONNET_4_6_CONFIG = {
@@ -147,6 +167,8 @@ export const CLAUDE_SONNET_4_6_CONFIG = {
gemini: 'gemini-2.0-flash', gemini: 'gemini-2.0-flash',
github: 'github:copilot', github: 'github:copilot',
codex: 'gpt-5.4', codex: 'gpt-5.4',
'nvidia-nim': 'nvidia/llama-3.1-nemotron-70b-instruct',
minimax: 'MiniMax-M2.5',
} as const satisfies ModelConfig } as const satisfies ModelConfig
// @[MODEL LAUNCH]: Register the new config here. // @[MODEL LAUNCH]: Register the new config here.

46
src/utils/model/minimaxModels.ts Normal file
View File

@@ -0,0 +1,46 @@
/**
* MiniMax model list for the /model picker.
* Full model catalog from MiniMax API.
*/
import type { ModelOption } from './modelOptions.js'
import { getAPIProvider } from './providers.js'
import { isEnvTruthy } from '../envUtils.js'
export function isMiniMaxProvider(): boolean {
if (isEnvTruthy(process.env.MINIMAX_API_KEY)) {
return true
}
const baseUrl = process.env.OPENAI_BASE_URL ?? ''
if (baseUrl.includes('minimax')) {
return true
}
return getAPIProvider() === 'minimax'
}
function getMiniMaxModels(): ModelOption[] {
return [
// Latest Generation Models - use correct MiniMax naming with M prefix
{ value: 'MiniMax-M2', label: 'MiniMax M2', description: 'MoE model - 131K context - Chat/Code/Reasoning' },
{ value: 'MiniMax-M2.1', label: 'MiniMax M2.1', description: 'Enhanced - 200K context - Vision' },
{ value: 'MiniMax-M2.5', label: 'MiniMax M2.5', description: 'Flagship - 256K context - Vision/Function-calling' },
{ value: 'MiniMax-Text-01', label: 'MiniMax Text 01', description: 'Text-focused - 512K context - FREE' },
{ value: 'MiniMax-Text-01-Preview', label: 'MiniMax Text 01 Preview', description: 'Preview - 256K context - FREE' },
{ value: 'MiniMax-Vision-01', label: 'MiniMax Vision 01', description: 'Vision model - 32K context' },
{ value: 'MiniMax-Vision-01-Fast', label: 'MiniMax Vision 01 Fast', description: 'Fast vision - 16K context - FREE' },
// Legacy free tier models
{ value: 'abab6.5s-chat', label: 'ABAB 6.5S Chat', description: 'Legacy free - 16K context' },
{ value: 'abab6.5-chat', label: 'ABAB 6.5 Chat', description: 'Legacy free - 32K context' },
{ value: 'abab6.5g-chat', label: 'ABAB 6.5G Chat', description: 'Generation 6.5 - 32K context' },
{ value: 'abab6-chat', label: 'ABAB 6 Chat', description: 'Legacy - 8K context' },
]
}
let cachedMiniMaxOptions: ModelOption[] | null = null
export function getCachedMiniMaxModelOptions(): ModelOption[] {
if (!cachedMiniMaxOptions) {
cachedMiniMaxOptions = getMiniMaxModels()
}
return cachedMiniMaxOptions
}

22
src/utils/model/modelOptions.ts
View File

@@ -35,6 +35,8 @@ import { has1mContext } from '../context.js'
import { getGlobalConfig } from '../config.js' import { getGlobalConfig } from '../config.js'
import { getActiveOpenAIModelOptionsCache } from '../providerProfiles.js' import { getActiveOpenAIModelOptionsCache } from '../providerProfiles.js'
import { getCachedOllamaModelOptions, isOllamaProvider } from './ollamaModels.js' import { getCachedOllamaModelOptions, isOllamaProvider } from './ollamaModels.js'
import { getCachedNvidiaNimModelOptions, isNvidiaNimProvider } from './nvidiaNimModels.js'
import { getCachedMiniMaxModelOptions, isMiniMaxProvider } from './minimaxModels.js'
import { getAntModels } from './antModels.js' import { getAntModels } from './antModels.js'
// @[MODEL LAUNCH]: Update all the available and default model option strings below. // @[MODEL LAUNCH]: Update all the available and default model option strings below.
@@ -390,6 +392,26 @@ function getModelOptionsBase(fastMode = false): ModelOption[] {
return [defaultOption] return [defaultOption]
} }
// When using NVIDIA NIM, show models from the NVIDIA catalog
if (isNvidiaNimProvider()) {
const defaultOption = getDefaultOptionForUser(fastMode)
const nvidiaModels = getCachedNvidiaNimModelOptions()
if (nvidiaModels.length > 0) {
return [defaultOption, ...nvidiaModels]
}
return [defaultOption]
}
// When using MiniMax, show models from the MiniMax catalog
if (isMiniMaxProvider()) {
const defaultOption = getDefaultOptionForUser(fastMode)
const minimaxModels = getCachedMiniMaxModelOptions()
if (minimaxModels.length > 0) {
return [defaultOption, ...minimaxModels]
}
return [defaultOption]
}
if (process.env.USER_TYPE === 'ant') { if (process.env.USER_TYPE === 'ant') {
// Build options from antModels config // Build options from antModels config
const antModelOptions: ModelOption[] = getAntModels().map(m => ({ const antModelOptions: ModelOption[] = getAntModels().map(m => ({

161
src/utils/model/nvidiaNimModels.ts Normal file
View File

@@ -0,0 +1,161 @@
/**
* NVIDIA NIM model list for the /model picker.
* Filtered to chat/instruct models only - embedding, reward, safety, vision, etc. excluded.
*/
import type { ModelOption } from './modelOptions.js'
import { getAPIProvider } from './providers.js'
import { isEnvTruthy } from '../envUtils.js'
export function isNvidiaNimProvider(): boolean {
// Check if explicitly set via NVIDIA_NIM or via provider flag
if (isEnvTruthy(process.env.NVIDIA_NIM)) {
return true
}
// Also check if using NVIDIA NIM endpoint
const baseUrl = process.env.OPENAI_BASE_URL ?? ''
if (baseUrl.includes('nvidia') || baseUrl.includes('integrate.api.nvidia')) {
return true
}
return getAPIProvider() === 'nvidia-nim'
}
function getNvidiaNimModels(): ModelOption[] {
return [
// AGENTIC REASONING MODELS
{ value: 'nvidia/cosmos-reason2-8b', label: 'Cosmos Reason 2 8B', description: 'Reasoning' },
{ value: 'microsoft/phi-4-mini-flash-reasoning', label: 'Phi 4 Mini Flash Reasoning', description: 'Reasoning' },
{ value: 'qwen/qwen3-next-80b-a3b-thinking', label: 'Qwen 3 Next 80B Thinking', description: 'Reasoning' },
{ value: 'deepseek-ai/deepseek-r1-distill-qwen-32b', label: 'DeepSeek R1 Qwen 32B', description: 'Reasoning' },
{ value: 'deepseek-ai/deepseek-r1-distill-qwen-14b', label: 'DeepSeek R1 Qwen 14B', description: 'Reasoning' },
{ value: 'deepseek-ai/deepseek-r1-distill-qwen-7b', label: 'DeepSeek R1 Qwen 7B', description: 'Reasoning' },
{ value: 'deepseek-ai/deepseek-r1-distill-llama-8b', label: 'DeepSeek R1 Llama 8B', description: 'Reasoning' },
{ value: 'qwen/qwq-32b', label: 'QwQ 32B Reasoning', description: 'Reasoning' },
// CODE MODELS
{ value: 'meta/codellama-70b', label: 'CodeLlama 70B', description: 'Code' },
{ value: 'bigcode/starcoder2-15b', label: 'StarCoder2 15B', description: 'Code' },
{ value: 'bigcode/starcoder2-7b', label: 'StarCoder2 7B', description: 'Code' },
{ value: 'mistralai/codestral-22b-instruct-v0.1', label: 'Codestral 22B', description: 'Code' },
{ value: 'mistralai/mamba-codestral-7b-v0.1', label: 'Mamba Codestral 7B', description: 'Code' },
{ value: 'deepseek-ai/deepseek-coder-6.7b-instruct', label: 'DeepSeek Coder 6.7B', description: 'Code' },
{ value: 'google/codegemma-7b', label: 'CodeGemma 7B', description: 'Code' },
{ value: 'google/codegemma-1.1-7b', label: 'CodeGemma 1.1 7B', description: 'Code' },
{ value: 'qwen/qwen2.5-coder-32b-instruct', label: 'Qwen 2.5 Coder 32B', description: 'Code' },
{ value: 'qwen/qwen2.5-coder-7b-instruct', label: 'Qwen 2.5 Coder 7B', description: 'Code' },
{ value: 'qwen/qwen3-coder-480b-a35b-instruct', label: 'Qwen 3 Coder 480B', description: 'Code' },
{ value: 'ibm/granite-34b-code-instruct', label: 'Granite 34B Code', description: 'Code' },
{ value: 'ibm/granite-8b-code-instruct', label: 'Granite 8B Code', description: 'Code' },
// NEMOTRON MODELS - NVIDIA Flagship
{ value: 'nvidia/llama-3.1-nemotron-70b-instruct', label: 'Nemotron 70B Instruct', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.1-nemotron-51b-instruct', label: 'Nemotron 51B Instruct', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.1-nemotron-ultra-253b-v1', label: 'Nemotron Ultra 253B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.3-nemotron-super-49b-v1', label: 'Nemotron Super 49B v1', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.3-nemotron-super-49b-v1.5', label: 'Nemotron Super 49B v1.5', description: 'NVIDIA Flagship' },
{ value: 'nvidia/nemotron-4-340b-instruct', label: 'Nemotron 4 340B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/nemotron-3-super-120b-a12b', label: 'Nemotron 3 Super 120B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/nemotron-3-nano-30b-a3b', label: 'Nemotron 3 Nano 30B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/nemotron-mini-4b-instruct', label: 'Nemotron Mini 4B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.1-nemotron-nano-8b-v1', label: 'Nemotron Nano 8B', description: 'NVIDIA Flagship' },
{ value: 'nvidia/llama-3.1-nemotron-nano-4b-v1.1', label: 'Nemotron Nano 4B v1.1', description: 'NVIDIA Flagship' },
// CHATQA MODELS
{ value: 'nvidia/llama3-chatqa-1.5-70b', label: 'Llama3 ChatQA 1.5 70B', description: 'Chat' },
{ value: 'nvidia/llama3-chatqa-1.5-8b', label: 'Llama3 ChatQA 1.5 8B', description: 'Chat' },
// META LLAMA MODELS
{ value: 'meta/llama-3.1-405b-instruct', label: 'Llama 3.1 405B', description: 'Meta Llama' },
{ value: 'meta/llama-3.1-70b-instruct', label: 'Llama 3.1 70B', description: 'Meta Llama' },
{ value: 'meta/llama-3.1-8b-instruct', label: 'Llama 3.1 8B', description: 'Meta Llama' },
{ value: 'meta/llama-3.2-90b-vision-instruct', label: 'Llama 3.2 90B Vision', description: 'Meta Llama' },
{ value: 'meta/llama-3.2-11b-vision-instruct', label: 'Llama 3.2 11B Vision', description: 'Meta Llama' },
{ value: 'meta/llama-3.2-3b-instruct', label: 'Llama 3.2 3B', description: 'Meta Llama' },
{ value: 'meta/llama-3.2-1b-instruct', label: 'Llama 3.2 1B', description: 'Meta Llama' },
{ value: 'meta/llama-3.3-70b-instruct', label: 'Llama 3.3 70B', description: 'Meta Llama' },
{ value: 'meta/llama-4-maverick-17b-128e-instruct', label: 'Llama 4 Maverick 17B', description: 'Meta Llama' },
{ value: 'meta/llama-4-scout-17b-16e-instruct', label: 'Llama 4 Scout 17B', description: 'Meta Llama' },
// GOOGLE GEMMA MODELS (text only - no vision)
{ value: 'google/gemma-4-31b-it', label: 'Gemma 4 31B', description: 'Google Gemma' },
{ value: 'google/gemma-3-27b-it', label: 'Gemma 3 27B', description: 'Google Gemma' },
{ value: 'google/gemma-3-12b-it', label: 'Gemma 3 12B', description: 'Google Gemma' },
{ value: 'google/gemma-3-4b-it', label: 'Gemma 3 4B', description: 'Google Gemma' },
{ value: 'google/gemma-3-1b-it', label: 'Gemma 3 1B', description: 'Google Gemma' },
{ value: 'google/gemma-3n-e4b-it', label: 'Gemma 3N E4B', description: 'Google Gemma' },
{ value: 'google/gemma-3n-e2b-it', label: 'Gemma 3N E2B', description: 'Google Gemma' },
{ value: 'google/gemma-2-27b-it', label: 'Gemma 2 27B', description: 'Google Gemma' },
{ value: 'google/gemma-2-9b-it', label: 'Gemma 2 9B', description: 'Google Gemma' },
{ value: 'google/gemma-2-2b-it', label: 'Gemma 2 2B', description: 'Google Gemma' },
// MISTRAL MODELS
{ value: 'mistralai/mistral-large-3-675b-instruct-2512', label: 'Mistral Large 3 675B', description: 'Mistral' },
{ value: 'mistralai/mistral-large-2-instruct', label: 'Mistral Large 2', description: 'Mistral' },
{ value: 'mistralai/mistral-large', label: 'Mistral Large', description: 'Mistral' },
{ value: 'mistralai/mistral-medium-3-instruct', label: 'Mistral Medium 3', description: 'Mistral' },
{ value: 'mistralai/mistral-small-4-119b-2603', label: 'Mistral Small 4 119B', description: 'Mistral' },
{ value: 'mistralai/mistral-small-3.1-24b-instruct-2503', label: 'Mistral Small 3.1 24B', description: 'Mistral' },
{ value: 'mistralai/mistral-small-24b-instruct', label: 'Mistral Small 24B', description: 'Mistral' },
{ value: 'mistralai/mistral-7b-instruct-v0.3', label: 'Mistral 7B v0.3', description: 'Mistral' },
{ value: 'mistralai/mistral-7b-instruct-v0.2', label: 'Mistral 7B v0.2', description: 'Mistral' },
{ value: 'mistralai/mixtral-8x22b-instruct-v0.1', label: 'Mixtral 8x22B', description: 'Mistral' },
{ value: 'mistralai/mixtral-8x22b-instruct-v0.1', label: 'Mixtral 8x22B Instruct', description: 'Mistral' },
{ value: 'mistralai/mixtral-8x7b-instruct-v0.1', label: 'Mixtral 8x7B', description: 'Mistral' },
{ value: 'mistralai/mistral-nemotron', label: 'Mistral Nemotron', description: 'Mistral' },
{ value: 'mistralai/mathstral-7b-v0.1', label: 'Mathstral 7B', description: 'Math' },
{ value: 'mistralai/ministral-14b-instruct-2512', label: 'Ministral 14B', description: 'Mistral' },
{ value: 'mistralai/devstral-2-123b-instruct-2512', label: 'Devstral 2 123B', description: 'Code' },
{ value: 'mistralai/magistral-small-2506', label: 'Magistral Small', description: 'Mistral' },
// MICROSOFT PHI MODELS (text only - no vision)
{ value: 'microsoft/phi-4-multimodal-instruct', label: 'Phi 4 Multimodal', description: 'Multimodal' },
{ value: 'microsoft/phi-4-mini-instruct', label: 'Phi 4 Mini', description: 'Phi' },
{ value: 'microsoft/phi-3.5-mini-instruct', label: 'Phi 3.5 Mini', description: 'Phi' },
{ value: 'microsoft/phi-3-small-128k-instruct', label: 'Phi 3 Small 128K', description: 'Phi' },
{ value: 'microsoft/phi-3-small-8k-instruct', label: 'Phi 3 Small 8K', description: 'Phi' },
{ value: 'microsoft/phi-3-medium-128k-instruct', label: 'Phi 3 Medium 128K', description: 'Phi' },
{ value: 'microsoft/phi-3-medium-4k-instruct', label: 'Phi 3 Medium 4K', description: 'Phi' },
{ value: 'microsoft/phi-3-mini-128k-instruct', label: 'Phi 3 Mini 128K', description: 'Phi' },
{ value: 'microsoft/phi-3-mini-4k-instruct', label: 'Phi 3 Mini 4K', description: 'Phi' },
// QWEN MODELS
{ value: 'qwen/qwen3.5-397b-a17b', label: 'Qwen 3.5 397B', description: 'Qwen' },
{ value: 'qwen/qwen3.5-122b-a10b', label: 'Qwen 3.5 122B', description: 'Qwen' },
{ value: 'qwen/qwen3-next-80b-a3b-instruct', label: 'Qwen 3 Next 80B', description: 'Qwen' },
{ value: 'qwen/qwen2.5-7b-instruct', label: 'Qwen 2.5 7B', description: 'Qwen' },
{ value: 'qwen/qwen2-7b-instruct', label: 'Qwen 2 7B', description: 'Qwen' },
{ value: 'qwen/qwen3-32b', label: 'Qwen 3 32B', description: 'Qwen' },
{ value: 'qwen/qwen3-8b', label: 'Qwen 3 8B', description: 'Qwen' },
// DEEPSEEK MODELS
{ value: 'deepseek-ai/deepseek-r1', label: 'DeepSeek R1', description: 'DeepSeek' },
{ value: 'deepseek-ai/deepseek-v3', label: 'DeepSeek V3', description: 'DeepSeek' },
{ value: 'deepseek-ai/deepseek-v3.2', label: 'DeepSeek V3.2', description: 'DeepSeek' },
{ value: 'deepseek-ai/deepseek-v3.1-terminus', label: 'DeepSeek V3.1 Terminus', description: 'DeepSeek' },
{ value: 'deepseek-ai/deepseek-v3.1', label: 'DeepSeek V3.1', description: 'DeepSeek' },
// IBM GRANITE MODELS
{ value: 'ibm/granite-3.3-8b-instruct', label: 'Granite 3.3 8B', description: 'IBM Granite' },
{ value: 'ibm/granite-3.0-8b-instruct', label: 'Granite 3.0 8B', description: 'IBM Granite' },
{ value: 'ibm/granite-3.0-3b-a800m-instruct', label: 'Granite 3.0 3B', description: 'IBM Granite' },
// OTHER MODELS
{ value: 'databricks/dbrx-instruct', label: 'DBRX Instruct', description: 'Other' },
{ value: '01-ai/yi-large', label: 'Yi Large', description: 'Other' },
{ value: 'ai21labs/jamba-1.5-large-instruct', label: 'Jamba 1.5 Large', description: 'Other' },
{ value: 'ai21labs/jamba-1.5-mini-instruct', label: 'Jamba 1.5 Mini', description: 'Other' },
{ value: 'writer/palmyra-creative-122b', label: 'Palmyra Creative 122B', description: 'Other' },
{ value: 'writer/palmyra-fin-70b-32k', label: 'Palmyra Fin 70B 32K', description: 'Other' },
{ value: 'writer/palmyra-med-70b', label: 'Palmyra Med 70B', description: 'Other' },
{ value: 'writer/palmyra-med-70b-32k', label: 'Palmyra Med 70B 32K', description: 'Other' },
// Z-AI GLM MODELS
{ value: 'z-ai/glm5', label: 'GLM-5', description: 'Z-AI' },
{ value: 'z-ai/glm4.7', label: 'GLM-4.7', description: 'Z-AI' },
// MINIMAX MODELS
{ value: 'minimaxai/minimax-m2.5', label: 'MiniMax M2.5', description: 'MiniMax' },
// MOONSHOT KIMI MODELS
{ value: 'moonshotai/kimi-k2.5', label: 'Kimi K2.5', description: 'Moonshot' },
{ value: 'moonshotai/kimi-k2-instruct', label: 'Kimi K2 Instruct', description: 'Moonshot' },
{ value: 'moonshotai/kimi-k2-thinking', label: 'Kimi K2 Thinking', description: 'Moonshot' },
{ value: 'moonshotai/kimi-k2.5-thinking', label: 'Kimi K2.5 Thinking', description: 'Moonshot' },
{ value: 'moonshotai/kimi-k2-instruct-0905', label: 'Kimi K2 Instruct 0905', description: 'Moonshot' },
]
}
let cachedNvidiaNimOptions: ModelOption[] | null = null
export function getCachedNvidiaNimModelOptions(): ModelOption[] {
if (!cachedNvidiaNimOptions) {
cachedNvidiaNimOptions = getNvidiaNimModels()
}
return cachedNvidiaNimOptions
}

95
src/utils/model/openaiContextWindows.ts
View File

@@ -104,6 +104,57 @@ const OPENAI_CONTEXT_WINDOWS: Record<string, number> = {
'devstral-latest': 256_000, 'devstral-latest': 256_000,
'ministral-3b-latest': 256_000, 'ministral-3b-latest': 256_000,
// NVIDIA NIM - popular models
'nvidia/llama-3.1-nemotron-70b-instruct': 128_000,
'nvidia/llama-3.1-nemotron-ultra-253b-v1': 128_000,
'nvidia/nemotron-mini-4b-instruct': 32_768,
'meta/llama-3.1-405b-instruct': 128_000,
'meta/llama-3.1-70b-instruct': 128_000,
'meta/llama-3.1-8b-instruct': 128_000,
'meta/llama-3.2-90b-instruct': 128_000,
'meta/llama-3.2-1b-instruct': 128_000,
'meta/llama-3.2-3b-instruct': 128_000,
'meta/llama-3.3-70b-instruct': 128_000,
// Google Gemma via NVIDIA NIM
'google/gemma-2-27b-it': 8_192,
'google/gemma-2-9b-it': 8_192,
'google/gemma-3-27b-it': 131_072,
'google/gemma-3-12b-it': 131_072,
'google/gemma-3-4b-it': 131_072,
// DeepSeek via NVIDIA NIM
'deepseek-ai/deepseek-r1': 128_000,
'deepseek-ai/deepseek-v3': 128_000,
'deepseek-ai/deepseek-v3.2': 128_000,
// Qwen via NVIDIA NIM
'qwen/qwen3-32b': 128_000,
'qwen/qwen3-8b': 128_000,
'qwen/qwen2.5-7b-instruct': 32_768,
// Mistral via NVIDIA NIM
'mistralai/mistral-large-3-675b-instruct-2512': 256_000,
'mistralai/mistral-large-2-instruct': 256_000,
'mistralai/mistral-small-3.1-24b-instruct-2503': 32_768,
'mistralai/mixtral-8x7b-instruct-v0.1': 32_768,
// Microsoft Phi via NVIDIA NIM
'microsoft/phi-4-mini-instruct': 16_384,
'microsoft/phi-3.5-mini-instruct': 16_384,
'microsoft/phi-3-mini-128k-instruct': 128_000,
// IBM Granite via NVIDIA NIM
'ibm/granite-3.3-8b-instruct': 8_192,
'ibm/granite-8b-code-instruct': 8_192,
// GLM models via NVIDIA NIM
'z-ai/glm5': 200_000,
'z-ai/glm4.7': 128_000,
// Kimi models via NVIDIA NIM
'moonshotai/kimi-k2.5': 200_000,
'moonshotai/kimi-k2-instruct': 128_000,
// DBRX via NVIDIA NIM
'databricks/dbrx-instruct': 131_072,
// Jamba via NVIDIA NIM
'ai21labs/jamba-1.5-large-instruct': 256_000,
'ai21labs/jamba-1.5-mini-instruct': 256_000,
// Yi via NVIDIA NIM
'01-ai/yi-large': 32_768,
// MiniMax (all M2.x variants share 204,800 context, 131,072 max output) // MiniMax (all M2.x variants share 204,800 context, 131,072 max output)
'MiniMax-M2.7': 204_800, 'MiniMax-M2.7': 204_800,
'MiniMax-M2.7-highspeed': 204_800, 'MiniMax-M2.7-highspeed': 204_800,
@@ -118,6 +169,13 @@ const OPENAI_CONTEXT_WINDOWS: Record<string, number> = {
'minimax-m2.1': 204_800, 'minimax-m2.1': 204_800,
'minimax-m2.1-highspeed': 204_800, 'minimax-m2.1-highspeed': 204_800,
// MiniMax new models
'MiniMax-Text-01': 524_288,
'MiniMax-Text-01-Preview': 262_144,
'MiniMax-Vision-01': 32_768,
'MiniMax-Vision-01-Fast': 16_384,
'MiniMax-M2': 204_800,
// Google (via OpenRouter) // Google (via OpenRouter)
'google/gemini-2.0-flash':1_048_576, 'google/gemini-2.0-flash':1_048_576,
'google/gemini-2.5-pro': 1_048_576, 'google/gemini-2.5-pro': 1_048_576,
@@ -246,6 +304,12 @@ const OPENAI_MAX_OUTPUT_TOKENS: Record<string, number> = {
'minimax-m2.5-highspeed': 131_072, 'minimax-m2.5-highspeed': 131_072,
'minimax-m2.1': 131_072, 'minimax-m2.1': 131_072,
'minimax-m2.1-highspeed': 131_072, 'minimax-m2.1-highspeed': 131_072,
// New MiniMax models
'MiniMax-M2': 131_072,
'MiniMax-Text-01': 65_536,
'MiniMax-Text-01-Preview': 65_536,
'MiniMax-Vision-01': 16_384,
'MiniMax-Vision-01-Fast': 16_384,
// Google (via OpenRouter) // Google (via OpenRouter)
'google/gemini-2.0-flash': 8_192, 'google/gemini-2.0-flash': 8_192,
@@ -266,11 +330,32 @@ const OPENAI_MAX_OUTPUT_TOKENS: Record<string, number> = {
'deepseek-r1:14b': 8_192, 'deepseek-r1:14b': 8_192,
'mistral:7b': 4_096, 'mistral:7b': 4_096,
'phi4:14b': 4_096, 'phi4:14b': 4_096,
'gemma2:27b': 4_096,
'codellama:13b': 4_096, // NVIDIA NIM models
'llama3.2:1b': 4_096, 'nvidia/llama-3.1-nemotron-70b-instruct': 32_768,
'qwen3:8b': 8_192, 'nvidia/nemotron-mini-4b-instruct': 8_192,
'codestral': 8_192, 'meta/llama-3.1-405b-instruct': 32_768,
'meta/llama-3.1-70b-instruct': 32_768,
'meta/llama-3.2-90b-instruct': 32_768,
'meta/llama-3.3-70b-instruct': 32_768,
'google/gemma-2-27b-it': 4_096,
'google/gemma-3-27b-it': 16_384,
'google/gemma-3-12b-it': 16_384,
'deepseek-ai/deepseek-r1': 32_768,
'deepseek-ai/deepseek-v3': 32_768,
'deepseek-ai/deepseek-v3.2': 32_768,
'qwen/qwen3-32b': 32_768,
'qwen/qwen2.5-7b-instruct': 8_192,
'mistralai/mistral-large-3-675b-instruct-2512': 32_768,
'mistralai/mixtral-8x7b-instruct-v0.1': 8_192,
'microsoft/phi-4-mini-instruct': 4_096,
'microsoft/phi-3.5-mini-instruct': 4_096,
'ibm/granite-3.3-8b-instruct': 4_096,
'z-ai/glm5': 32_768,
'moonshotai/kimi-k2.5': 32_768,
'databricks/dbrx-instruct': 32_768,
'ai21labs/jamba-1.5-large-instruct': 32_768,
'01-ai/yi-large': 8_192,
} }
function lookupByModel<T>(table: Record<string, T>, model: string): T | undefined { function lookupByModel<T>(table: Record<string, T>, model: string): T | undefined {

8
src/utils/model/providers.ts
View File

@@ -11,9 +11,17 @@ export type APIProvider =
| 'gemini' | 'gemini'
| 'github' | 'github'
| 'codex' | 'codex'
| 'nvidia-nim'
| 'minimax'
| 'mistral' | 'mistral'
export function getAPIProvider(): APIProvider { export function getAPIProvider(): APIProvider {
if (isEnvTruthy(process.env.NVIDIA_NIM)) {
return 'nvidia-nim'
}
if (isEnvTruthy(process.env.MINIMAX_API_KEY)) {
return 'minimax'
}
return isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI) return isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI)
? 'gemini' ? 'gemini'
: :

36
src/utils/model/validateModel.ts
View File

@@ -11,6 +11,8 @@ import {
} from '@anthropic-ai/sdk' } from '@anthropic-ai/sdk'
import { getModelStrings } from './modelStrings.js' import { getModelStrings } from './modelStrings.js'
import { getCachedOllamaModelOptions, isOllamaProvider } from './ollamaModels.js' import { getCachedOllamaModelOptions, isOllamaProvider } from './ollamaModels.js'
import { getCachedNvidiaNimModelOptions, isNvidiaNimProvider } from './nvidiaNimModels.js'
import { getCachedMiniMaxModelOptions, isMiniMaxProvider } from './minimaxModels.js'
// Cache valid models to avoid repeated API calls // Cache valid models to avoid repeated API calls
const validModelCache = new Map<string, boolean>() const validModelCache = new Map<string, boolean>()
@@ -47,6 +49,40 @@ export async function validateModel(
// If cache is empty, fall through to API validation // If cache is empty, fall through to API validation
} }
// For NVIDIA NIM provider, validate against cached model list
if (isNvidiaNimProvider()) {
const nvidiaModels = getCachedNvidiaNimModelOptions()
const found = nvidiaModels.some(m => m.value === normalizedModel)
if (found) {
validModelCache.set(normalizedModel, true)
return { valid: true }
}
if (nvidiaModels.length > 0) {
const MAX_SHOWN = 5
const names = nvidiaModels.map(m => m.value)
const shown = names.slice(0, MAX_SHOWN).join(', ')
const suffix = names.length > MAX_SHOWN ? ` and ${names.length - MAX_SHOWN} more` : ''
return { valid: false, error: `Model '${normalizedModel}' not found in NVIDIA NIM catalog. Available: ${shown}${suffix}` }
}
}
// For MiniMax provider, validate against cached model list
if (isMiniMaxProvider()) {
const minimaxModels = getCachedMiniMaxModelOptions()
const found = minimaxModels.some(m => m.value === normalizedModel)
if (found) {
validModelCache.set(normalizedModel, true)
return { valid: true }
}
if (minimaxModels.length > 0) {
const MAX_SHOWN = 5
const names = minimaxModels.map(m => m.value)
const shown = names.slice(0, MAX_SHOWN).join(', ')
const suffix = names.length > MAX_SHOWN ? ` and ${names.length - MAX_SHOWN} more` : ''
return { valid: false, error: `Model '${normalizedModel}' not found in MiniMax catalog. Available: ${shown}${suffix}` }
}
}
// Check against availableModels allowlist before any API call // Check against availableModels allowlist before any API call
if (!isModelAllowed(normalizedModel)) { if (!isModelAllowed(normalizedModel)) {
return { return {

8
src/utils/providerDiscovery.ts
View File

@@ -105,6 +105,14 @@ export function getLocalOpenAICompatibleProviderLabel(baseUrl?: string): string
) { ) {
return 'text-generation-webui' return 'text-generation-webui'
} }
// Check for NVIDIA NIM
if (host.includes('nvidia') || haystack.includes('nvidia') || host.includes('integrate.api.nvidia')) {
return 'NVIDIA NIM'
}
// Check for MiniMax (both api.minimax.io and api.minimax.chat)
if (host.includes('minimax') || haystack.includes('minimax')) {
return 'MiniMax'
}
} catch { } catch {
// Fall back to the generic label when the base URL is malformed. // Fall back to the generic label when the base URL is malformed.
} }

17
src/utils/providerFlag.ts
View File

@@ -21,6 +21,8 @@ export const VALID_PROVIDERS = [
'bedrock', 'bedrock',
'vertex', 'vertex',
'ollama', 'ollama',
'nvidia-nim',
'minimax',
] as const ] as const
export type ProviderFlagName = (typeof VALID_PROVIDERS)[number] export type ProviderFlagName = (typeof VALID_PROVIDERS)[number]
@@ -131,6 +133,21 @@ export function applyProviderFlag(
} }
if (model) process.env.OPENAI_MODEL = model if (model) process.env.OPENAI_MODEL = model
break break
case 'nvidia-nim':
process.env.CLAUDE_CODE_USE_OPENAI = '1'
process.env.OPENAI_BASE_URL ??= 'https://integrate.api.nvidia.com/v1'
process.env.NVIDIA_NIM = '1'
process.env.OPENAI_MODEL ??= 'nvidia/llama-3.1-nemotron-70b-instruct'
if (model) process.env.OPENAI_MODEL = model
break
case 'minimax':
process.env.CLAUDE_CODE_USE_OPENAI = '1'
process.env.OPENAI_BASE_URL ??= 'https://api.minimax.io/v1'
process.env.OPENAI_MODEL ??= 'MiniMax-M2.5'
if (model) process.env.OPENAI_MODEL = model
break
} }
return {} return {}

175
src/utils/providerProfile.ts
View File

@@ -6,29 +6,29 @@ import {
isCodexBaseUrl, isCodexBaseUrl,
resolveCodexApiCredentials, resolveCodexApiCredentials,
resolveProviderRequest, resolveProviderRequest,
} from '../services/api/providerConfig.ts' } from '../services/api/providerConfig.js'
import { parseChatgptAccountId } from '../services/api/codexOAuthShared.js' import { parseChatgptAccountId } from '../services/api/codexOAuthShared.js'
import { import {
getGoalDefaultOpenAIModel, getGoalDefaultOpenAIModel,
normalizeRecommendationGoal, normalizeRecommendationGoal,
type RecommendationGoal, type RecommendationGoal,
} from './providerRecommendation.ts' } from './providerRecommendation.js'
import { readGeminiAccessToken } from './geminiCredentials.ts' import { readGeminiAccessToken } from './geminiCredentials.js'
import { getOllamaChatBaseUrl } from './providerDiscovery.ts' import { getOllamaChatBaseUrl } from './providerDiscovery.js'
import { getProviderValidationError } from './providerValidation.ts' import { getProviderValidationError } from './providerValidation.js'
import { import {
maskSecretForDisplay, maskSecretForDisplay,
redactSecretValueForDisplay, redactSecretValueForDisplay,
sanitizeApiKey, sanitizeApiKey,
sanitizeProviderConfigValue, sanitizeProviderConfigValue,
} from './providerSecrets.ts' } from './providerSecrets.js'
export { export {
maskSecretForDisplay, maskSecretForDisplay,
redactSecretValueForDisplay, redactSecretValueForDisplay,
sanitizeApiKey, sanitizeApiKey,
sanitizeProviderConfigValue, sanitizeProviderConfigValue,
} from './providerSecrets.ts' } from './providerSecrets.js'
export const PROFILE_FILE_NAME = '.openclaude-profile.json' export const PROFILE_FILE_NAME = '.openclaude-profile.json'
export const DEFAULT_GEMINI_BASE_URL = export const DEFAULT_GEMINI_BASE_URL =
@@ -57,18 +57,28 @@ const PROFILE_ENV_KEYS = [
'GEMINI_MODEL', 'GEMINI_MODEL',
'GEMINI_BASE_URL', 'GEMINI_BASE_URL',
'GOOGLE_API_KEY', 'GOOGLE_API_KEY',
'NVIDIA_NIM',
'NVIDIA_API_KEY',
'NVIDIA_MODEL',
'MINIMAX_API_KEY',
'MINIMAX_BASE_URL',
'MINIMAX_MODEL',
'MISTRAL_BASE_URL', 'MISTRAL_BASE_URL',
'MISTRAL_API_KEY', 'MISTRAL_API_KEY',
'MISTRAL_MODEL', 'MISTRAL_MODEL',
] as const ] as const
export type ProviderProfile = const SECRET_ENV_KEYS = [
| 'openai' 'OPENAI_API_KEY',
| 'ollama' 'CODEX_API_KEY',
| 'codex' 'GEMINI_API_KEY',
| 'gemini' 'GOOGLE_API_KEY',
| 'atomic-chat' 'NVIDIA_API_KEY',
| 'mistral' 'MINIMAX_API_KEY',
'MISTRAL_API_KEY',
] as const
export type ProviderProfile = 'openai' | 'ollama' | 'codex' | 'gemini' | 'atomic-chat' | 'nvidia-nim' | 'minimax' | 'mistral'
export type ProfileEnv = { export type ProfileEnv = {
OPENAI_BASE_URL?: string OPENAI_BASE_URL?: string
@@ -82,6 +92,12 @@ export type ProfileEnv = {
GEMINI_AUTH_MODE?: 'api-key' | 'access-token' | 'adc' GEMINI_AUTH_MODE?: 'api-key' | 'access-token' | 'adc'
GEMINI_MODEL?: string GEMINI_MODEL?: string
GEMINI_BASE_URL?: string GEMINI_BASE_URL?: string
GOOGLE_API_KEY?: string
NVIDIA_NIM?: string
NVIDIA_API_KEY?: string
MINIMAX_API_KEY?: string
MINIMAX_BASE_URL?: string
MINIMAX_MODEL?: string
MISTRAL_BASE_URL?: string MISTRAL_BASE_URL?: string
MISTRAL_API_KEY?: string MISTRAL_API_KEY?: string
MISTRAL_MODEL?: string MISTRAL_MODEL?: string
@@ -93,6 +109,19 @@ export type ProfileFile = {
createdAt: string createdAt: string
} }
type SecretValueSource = Partial<
Record<
| 'OPENAI_API_KEY'
| 'CODEX_API_KEY'
| 'GEMINI_API_KEY'
| 'GOOGLE_API_KEY'
| 'NVIDIA_API_KEY'
| 'MINIMAX_API_KEY'
| 'MISTRAL_API_KEY',
string | undefined
>
>
type ProfileFileLocation = { type ProfileFileLocation = {
cwd?: string cwd?: string
filePath?: string filePath?: string
@@ -113,6 +142,8 @@ export function isProviderProfile(value: unknown): value is ProviderProfile {
value === 'codex' || value === 'codex' ||
value === 'gemini' || value === 'gemini' ||
value === 'atomic-chat' || value === 'atomic-chat' ||
value === 'nvidia-nim' ||
value === 'minimax' ||
value === 'mistral' value === 'mistral'
) )
} }
@@ -143,6 +174,67 @@ export function buildAtomicChatProfileEnv(
} }
} }
export function buildNvidiaNimProfileEnv(options: {
model?: string | null
baseUrl?: string | null
apiKey?: string | null
processEnv?: NodeJS.ProcessEnv
}): ProfileEnv | null {
const processEnv = options.processEnv ?? process.env
const key = sanitizeApiKey(options.apiKey ?? processEnv.NVIDIA_API_KEY)
if (!key) {
return null
}
const defaultBaseUrl = 'https://integrate.api.nvidia.com/v1'
const secretSource: SecretValueSource = { OPENAI_API_KEY: key }
return {
OPENAI_BASE_URL:
sanitizeProviderConfigValue(options.baseUrl, secretSource) ||
sanitizeProviderConfigValue(processEnv.OPENAI_BASE_URL, secretSource) ||
defaultBaseUrl,
OPENAI_MODEL:
sanitizeProviderConfigValue(options.model, secretSource) ||
sanitizeProviderConfigValue(processEnv.OPENAI_MODEL, secretSource) ||
'nvidia/llama-3.1-nemotron-70b-instruct',
OPENAI_API_KEY: key,
NVIDIA_NIM: '1',
}
}
export function buildMiniMaxProfileEnv(options: {
model?: string | null
baseUrl?: string | null
apiKey?: string | null
processEnv?: NodeJS.ProcessEnv
}): ProfileEnv | null {
const processEnv = options.processEnv ?? process.env
const key = sanitizeApiKey(options.apiKey ?? processEnv.MINIMAX_API_KEY)
if (!key) {
return null
}
const defaultBaseUrl = 'https://api.minimax.io/v1'
const defaultModel = 'MiniMax-M2.5'
const secretSource: SecretValueSource = { OPENAI_API_KEY: key }
return {
OPENAI_BASE_URL:
sanitizeProviderConfigValue(options.baseUrl, secretSource) ||
sanitizeProviderConfigValue(processEnv.OPENAI_BASE_URL, secretSource) ||
defaultBaseUrl,
OPENAI_MODEL:
sanitizeProviderConfigValue(options.model, secretSource) ||
sanitizeProviderConfigValue(processEnv.OPENAI_MODEL, secretSource) ||
defaultModel,
OPENAI_API_KEY: key,
MINIMAX_API_KEY: key,
MINIMAX_BASE_URL: defaultBaseUrl,
MINIMAX_MODEL: defaultModel,
}
}
export function buildGeminiProfileEnv(options: { export function buildGeminiProfileEnv(options: {
model?: string | null model?: string | null
baseUrl?: string | null baseUrl?: string | null
@@ -161,15 +253,13 @@ export function buildGeminiProfileEnv(options: {
return null return null
} }
const secretSource: SecretValueSource = key ? { GEMINI_API_KEY: key } : {}
const env: ProfileEnv = { const env: ProfileEnv = {
GEMINI_AUTH_MODE: authMode, GEMINI_AUTH_MODE: authMode,
GEMINI_MODEL: GEMINI_MODEL:
sanitizeProviderConfigValue(options.model, { GEMINI_API_KEY: key }, processEnv) || sanitizeProviderConfigValue(options.model, secretSource) ||
sanitizeProviderConfigValue( sanitizeProviderConfigValue(processEnv.GEMINI_MODEL, secretSource) ||
processEnv.GEMINI_MODEL,
{ GEMINI_API_KEY: key },
processEnv,
) ||
DEFAULT_GEMINI_MODEL, DEFAULT_GEMINI_MODEL,
} }
@@ -178,12 +268,8 @@ export function buildGeminiProfileEnv(options: {
} }
const baseUrl = const baseUrl =
sanitizeProviderConfigValue(options.baseUrl, { GEMINI_API_KEY: key }, processEnv) || sanitizeProviderConfigValue(options.baseUrl, secretSource) ||
sanitizeProviderConfigValue( sanitizeProviderConfigValue(processEnv.GEMINI_BASE_URL, secretSource)
processEnv.GEMINI_BASE_URL,
{ GEMINI_API_KEY: key },
processEnv,
)
if (baseUrl) { if (baseUrl) {
env.GEMINI_BASE_URL = baseUrl env.GEMINI_BASE_URL = baseUrl
} }
@@ -205,15 +291,14 @@ export function buildOpenAIProfileEnv(options: {
} }
const defaultModel = getGoalDefaultOpenAIModel(options.goal) const defaultModel = getGoalDefaultOpenAIModel(options.goal)
const secretSource: SecretValueSource = { OPENAI_API_KEY: key }
const shellOpenAIModel = sanitizeProviderConfigValue( const shellOpenAIModel = sanitizeProviderConfigValue(
processEnv.OPENAI_MODEL, processEnv.OPENAI_MODEL,
{ OPENAI_API_KEY: key }, secretSource,
processEnv,
) )
const shellOpenAIBaseUrl = sanitizeProviderConfigValue( const shellOpenAIBaseUrl = sanitizeProviderConfigValue(
processEnv.OPENAI_BASE_URL, processEnv.OPENAI_BASE_URL,
{ OPENAI_API_KEY: key }, secretSource,
processEnv,
) )
const shellOpenAIRequest = resolveProviderRequest({ const shellOpenAIRequest = resolveProviderRequest({
model: shellOpenAIModel, model: shellOpenAIModel,
@@ -224,19 +309,11 @@ export function buildOpenAIProfileEnv(options: {
return { return {
OPENAI_BASE_URL: OPENAI_BASE_URL:
sanitizeProviderConfigValue( sanitizeProviderConfigValue(options.baseUrl, secretSource) ||
options.baseUrl,
{ OPENAI_API_KEY: key },
processEnv,
) ||
(useShellOpenAIConfig ? shellOpenAIBaseUrl : undefined) || (useShellOpenAIConfig ? shellOpenAIBaseUrl : undefined) ||
DEFAULT_OPENAI_BASE_URL, DEFAULT_OPENAI_BASE_URL,
OPENAI_MODEL: OPENAI_MODEL:
sanitizeProviderConfigValue( sanitizeProviderConfigValue(options.model, secretSource) ||
options.model,
{ OPENAI_API_KEY: key },
processEnv,
) ||
(useShellOpenAIConfig ? shellOpenAIModel : undefined) || (useShellOpenAIConfig ? shellOpenAIModel : undefined) ||
defaultModel, defaultModel,
OPENAI_API_KEY: key, OPENAI_API_KEY: key,
@@ -293,21 +370,19 @@ export function buildMistralProfileEnv(options: {
const env: ProfileEnv = { const env: ProfileEnv = {
MISTRAL_API_KEY: key, MISTRAL_API_KEY: key,
MISTRAL_MODEL: MISTRAL_MODEL:
sanitizeProviderConfigValue(options.model, { MISTRAL_API_KEY: key }, processEnv) || sanitizeProviderConfigValue(options.model, { MISTRAL_API_KEY: key }) ||
sanitizeProviderConfigValue( sanitizeProviderConfigValue(
processEnv.MISTRAL_MODEL, processEnv.MISTRAL_MODEL,
{ MISTRAL_API_KEY: key }, { MISTRAL_API_KEY: key },
processEnv,
) || ) ||
DEFAULT_MISTRAL_MODEL, DEFAULT_MISTRAL_MODEL,
} }
const baseUrl = const baseUrl =
sanitizeProviderConfigValue(options.baseUrl, { MISTRAL_API_KEY: key }, processEnv) || sanitizeProviderConfigValue(options.baseUrl, { MISTRAL_API_KEY: key }) ||
sanitizeProviderConfigValue( sanitizeProviderConfigValue(
processEnv.MISTRAL_BASE_URL, processEnv.MISTRAL_BASE_URL,
{ MISTRAL_API_KEY: key }, { MISTRAL_API_KEY: key },
processEnv,
) )
if (baseUrl) { if (baseUrl) {
env.MISTRAL_BASE_URL = baseUrl env.MISTRAL_BASE_URL = baseUrl
@@ -465,11 +540,11 @@ export async function buildLaunchEnv(options: {
) )
const shellOpenAIModel = sanitizeProviderConfigValue( const shellOpenAIModel = sanitizeProviderConfigValue(
processEnv.OPENAI_MODEL, processEnv.OPENAI_MODEL,
processEnv, processEnv as SecretValueSource,
) )
const shellOpenAIBaseUrl = sanitizeProviderConfigValue( const shellOpenAIBaseUrl = sanitizeProviderConfigValue(
processEnv.OPENAI_BASE_URL, processEnv.OPENAI_BASE_URL,
processEnv, processEnv as SecretValueSource,
) )
const persistedGeminiModel = sanitizeProviderConfigValue( const persistedGeminiModel = sanitizeProviderConfigValue(
persistedEnv.GEMINI_MODEL, persistedEnv.GEMINI_MODEL,
@@ -481,11 +556,11 @@ export async function buildLaunchEnv(options: {
) )
const shellGeminiModel = sanitizeProviderConfigValue( const shellGeminiModel = sanitizeProviderConfigValue(
processEnv.GEMINI_MODEL, processEnv.GEMINI_MODEL,
processEnv, processEnv as SecretValueSource,
) )
const shellGeminiBaseUrl = sanitizeProviderConfigValue( const shellGeminiBaseUrl = sanitizeProviderConfigValue(
processEnv.GEMINI_BASE_URL, processEnv.GEMINI_BASE_URL,
processEnv, processEnv as SecretValueSource,
) )
const shellGeminiAccessToken = const shellGeminiAccessToken =
processEnv.GEMINI_ACCESS_TOKEN?.trim() || undefined processEnv.GEMINI_ACCESS_TOKEN?.trim() || undefined
@@ -567,19 +642,15 @@ export async function buildLaunchEnv(options: {
const shellMistralModel = sanitizeProviderConfigValue( const shellMistralModel = sanitizeProviderConfigValue(
processEnv.MISTRAL_MODEL, processEnv.MISTRAL_MODEL,
processEnv,
) )
const persistedMistralModel = sanitizeProviderConfigValue( const persistedMistralModel = sanitizeProviderConfigValue(
persistedEnv.MISTRAL_MODEL, persistedEnv.MISTRAL_MODEL,
persistedEnv,
) )
const shellMistralBaseUrl = sanitizeProviderConfigValue( const shellMistralBaseUrl = sanitizeProviderConfigValue(
processEnv.MISTRAL_BASE_URL, processEnv.MISTRAL_BASE_URL,
processEnv,
) )
const persistedMistralBaseUrl = sanitizeProviderConfigValue( const persistedMistralBaseUrl = sanitizeProviderConfigValue(
persistedEnv.MISTRAL_BASE_URL, persistedEnv.MISTRAL_BASE_URL,
persistedEnv,
) )
env.MISTRAL_MODEL = env.MISTRAL_MODEL =

33
src/utils/providerProfiles.ts
View File

@@ -20,6 +20,8 @@ export type ProviderPreset =
| 'openrouter' | 'openrouter'
| 'lmstudio' | 'lmstudio'
| 'custom' | 'custom'
| 'nvidia-nim'
| 'minimax'
export type ProviderProfileInput = { export type ProviderProfileInput = {
provider?: ProviderProfile['provider'] provider?: ProviderProfile['provider']
@@ -229,6 +231,24 @@ export function getProviderPresetDefaults(
apiKey: process.env.OPENAI_API_KEY ?? '', apiKey: process.env.OPENAI_API_KEY ?? '',
requiresApiKey: false, requiresApiKey: false,
} }
case 'nvidia-nim':
return {
provider: 'openai',
name: 'NVIDIA NIM',
baseUrl: 'https://integrate.api.nvidia.com/v1',
model: 'nvidia/llama-3.1-nemotron-70b-instruct',
apiKey: process.env.NVIDIA_API_KEY ?? '',
requiresApiKey: true,
}
case 'minimax':
return {
provider: 'openai',
name: 'MiniMax',
baseUrl: 'https://api.minimax.io/v1',
model: 'MiniMax-M2.5',
apiKey: process.env.MINIMAX_API_KEY ?? '',
requiresApiKey: true,
}
case 'ollama': case 'ollama':
default: default:
return { return {
@@ -365,6 +385,11 @@ export function clearProviderProfileEnvFromProcessEnv(
delete processEnv.ANTHROPIC_API_KEY delete processEnv.ANTHROPIC_API_KEY
delete processEnv[PROFILE_ENV_APPLIED_FLAG] delete processEnv[PROFILE_ENV_APPLIED_FLAG]
delete processEnv[PROFILE_ENV_APPLIED_ID] delete processEnv[PROFILE_ENV_APPLIED_ID]
// Clear provider-specific API keys
delete processEnv.MINIMAX_API_KEY
delete processEnv.NVIDIA_API_KEY
delete processEnv.NVIDIA_NIM
} }
export function applyProviderProfileToProcessEnv(profile: ProviderProfile): void { export function applyProviderProfileToProcessEnv(profile: ProviderProfile): void {
@@ -395,6 +420,14 @@ export function applyProviderProfileToProcessEnv(profile: ProviderProfile): void
if (profile.apiKey) { if (profile.apiKey) {
process.env.OPENAI_API_KEY = profile.apiKey process.env.OPENAI_API_KEY = profile.apiKey
// Also set provider-specific API keys for detection
const baseUrl = profile.baseUrl.toLowerCase()
if (baseUrl.includes('minimax')) {
process.env.MINIMAX_API_KEY = profile.apiKey
}
if (baseUrl.includes('nvidia') || baseUrl.includes('integrate.api.nvidia')) {
process.env.NVIDIA_API_KEY = profile.apiKey
}
} else { } else {
delete process.env.OPENAI_API_KEY delete process.env.OPENAI_API_KEY
} }